MIT Student Builds Autonomous Self-Flying Drone with Obstacle Avoidance

MIT Autonomous Robotic Self Flying Drone Quadcopter

Andy Barry, an MIT grad student set off to build an autonomous drone that was not only capable of flying at speeds up to 30MPH, but it could also automatically avoid obstacles that stands in its way on-the-fly.

The drone itself isn’t all that special, its being powered by a processor found in the Samsung Galaxy S3 (quad-core CPU’s), weighs in at a light 664 grams, and it also has two stereo cameras onboard. The real brains behind this smart drone is the software that runs on it.

The processors compute the flight trajectory, where obstacles are, and dozens of other parameters for the flight within milliseconds as its flying through the air at speeds up to 30 MPH.

One of the most significant things about this project is that the drone has no idea what’s about to happen or what it will do next until just milliseconds before. It operates similar to how self-driving cars operate. If something out of the normal happens, such as a branch or tree gets in the way – it must make split second decisions. It’s always scanning about 10 meters ahead for obstacles, which gives it enough time to react.

The possibilities with this technology are endless and will eventually make its way into everyday hobbyist drones like the Q-Cop 450 we recently reviewed. Imagine your Q-Cop 450 flying along and you nearly take out a branch but instead the drone takes control and steers you away from danger and returns control back to you when things are safe again.

Thanks to Barry’s openness to help the drone community, he has shared his work with anyone that wants to review it on Github.

 

About The Author

James Anderson

James is the owner and editor-in-chief of Geek Inspector. He enjoys technology, motorcycles, traveling, and spending time with his family.

Related posts